methods of geological engineering in discontinuous rocks

ods of Geoogica

9 neer.ng in Discontin ous Rocks

RICHARD E. GOOD N Professor of Geological Engineering

University of California, Berkeley

WEST PUBLISHING COMPANY ST. PAUL NEW YORK BOSTON LOS ANGELES SAN FRANCISCO

COPYRIGHT 1976 By WEST PUBLISHING CO. All rights reserved Printed in the United States of America

Library of Congress Cataloging in Publication Data

Goodman, Richard E. Methods of geological engineering in discontinuous rocks. Includes index. 1. Rock mechanics. I. Title.

TA706.G66 624'.1513 75-42152

ISBN 0-8299-0066-7

preface

This work contains material from my courses at Berkeley in

Engineering Geology and Applied Rock Mechanics. It could serve as a

supplementary source for general courses in these fields, as well as

a text in a more specialized course in geological engineering analysis.

I have not tried to separate the disciplines of geology and civil

engineering as I think it unnatural to do so. The geologist can

decide on the relevancy of the features he maps and describes only

if he understands how they relate to the analytical process. The

engineer can not handle the geological data correctly without the

respect for its !! fuzziness II that comes from a personal acquain tance

with geology.

Interest in geological engineering methods has occupied me since

I visited the sites of the Malpasset and Vajont failures in the early

60 1 s. I wanted to know if a person with reasonable education and

experience in engineering geology could have foreseen the difficulties.

It seemed to me that objective, rational procedures for evaluating

such sites were inadequate and that we were, perhaps, relying too

v

vi Preface

heavily on intuition. Wisdom derived from real experiences will

always be an important and necessary ingredient for predicting rock

behavior; but we must also search for objective tools. Since 1960,

a number of such tools have appeared and it is timely and useful to

set them forth for students, and for practicing technicians who can

apply them in their work.

Most of this Monograph was written at Imperial College, London,

during the term of a Guggenheim Fellowship, while I was on a sabbatical

leave from Berkeley. I wish to thank the John Simon Guggenheim Founda-

tion and the Regents of the University of California for this opportu-

nity to reflect on the whole of the field of geological engineering

methodology. In residency at the Royal School of Mines during 1973, I

was privileged to frequent discussions with Professor Evert Hoek, Dr.

John Bray, John Boyd, and other faculty members. And I profited from

the work of a number of their students, past and contemporary, includ-

ing Peter Cundall, Christopher St. John, Nick Barton, Tidu Maini, John

Franklin, John Sharp, Laurie Richards, Dermot Ross-Brown, Peter Riley,

Ross Hammett, Peter Kelsall, Don Moy, Tim Harper, and Graeme Major. I

am particularly indebted to Dr. St. John who prepared the special

finite element program listed in Appendix 1. We were anxious to have

a small finite element program written expressly for a student trying

to bridge between theory and application.

In the text, I have indicated the sources for ideas by references

to the works of numerous authors. I have profited from personal com-

munications over the years with a number of these authors - - Walter

Wittke, Klaus John, Pierre Lande, and "Skip" Hendron with respect to

limit equilibrium analyses - - Dr. Leopold Muller and Karel Drozd with

Preface

respect to p h ysical models E. J. Polak, Tor Brekke, and Dan Moye

wi t h respect to geological and geophysical exploration - - and Robert

Tay l o r , Edward Wilson , Hugh Trollope, and Ann Bornstein with respect

to comp uter me t h o ds. I have also benefitted from the wo rk of past and

p res en t students at Berkeley inc l uding Yuzo Ohnishi, P. N. Sundaram,

Ashraf Mahtab, Rudolfo de la Cruz, John Cadman, Jacques Dubois, Alain

de Rouvray , and Francoi s Heuze, and of post Doctoral scholars Kemal

Erguvanl i, J ean Luc Dessenne, and Karel Drozd. The critical comments

and sugge s tio ns of Pr of . Arvid Johnson of Stanford University, and

Prof . Hendron and hi s colleagues at the University of Illinois were

quite h elpful .

A n umbe r of person s kindly loaned materials or gave me permission

to refer to t he ir work. These include Nick Barton, Z. T. Bieniawski,

Tor Brek k e , W. Chin n , Lloyd Cluff, Phillip Cole, James Coulson, Walter

Day , William De a r man, G. Everling, Irving Fatt, Alena Gralewska-Vickery,

Ri chard Ha y , Francois Heuze, Larry James, Dennis Lachel, Branko Ladanyi,

Thomas Lan g , Pierre Lon de, Ken Ma tthews, J. Myung, Carlos Ospina, Marc

Pane t, M. Pop ovic , Howard Pratt, Hernando Quijano, Doro t h y Radbruch,

Ni ck Ren ge r s , Manuel Ro cha, Fritz Rummel, F. Sabarly, Ed. Slebir,

Gera rdo Tarna, Jose Tejada , Ruth Terz aghi, Cl. Tourenq, Lloyd Underwood ,

a nd J oel Verdi e r.

The f ollowi n g o rgan izations generously permitted me to refer to

or borrow thei r mat e ri als: Atlas Copco ABEM ( S tockholm), Bergbau

Forschung (Essen) , Birdwe l l Division of Seismograph Service Corp.

(Tulsa ) , Californi a Department of Water Resources (Sacramento),

Christensen Diamond Produ c ts Co . (Salt Lake City), Coyne and Bellier

(Par is ) J Engineering Laboratory Equipment Ltd. (Hemel Hempstead, U. K.),

viii Preface

Golder and Brawner and Assoc. (Vancouver), Ingetec Ltda. (Bogota),

Integral Ltda. (Medellin), Laboratoire des Ponts et Chaussees (Paris),

Laboratorio Nacional de Engenharia Civil (Lisbon), Longyear Co.

(Minneapolis), Joy Manufacturing Co. (Montgomeryville, Pa.), Mindrill

Ltd. (Melbourne), Norwegian Geotechnical Institute (Oslo), Soil

Mechanics Equipment Co. (Glen Ellyn, Ill.), Sprague and Henwood, Inc.

(Scranton, Pa.), Tacoma City Light, TerraTek (Salt Lake City), U. S.

Army Corps of Engineers (Libby Resident Office; Missouri River

Division, Omaha District, and Explosives Excavation Research Lab.),

U. S. Bureau of Mines, (Denver), U. S. Bureau of Reclamation (Denver),

U. S. Geological Survey (Menlo Park), Woodward Clyde and Associates

(Oakland), and Zavod Za Geotehniku I Fundiranje (Sarajevo).

Finally, I wish to thank the persons who helped me with the work

of producing the manuscript: Fran Riley, Laurie Wilson, Gloria

Pelatowski, and Lillian Goodman.

to the memory of

Parker D. Trask

ix

contents

PREFACE

1. II"TRODUCTI ON

2. ROCK CLASSIF ICATI ON 14

The Nature of Rock 14

Rock Specimen Versus Rock Mass 15

Petrologic Classification of Rock Specimens 16

Rock Versus Soil and Weathered Rock 19

Weathering 22

Index Tests for the Quality of the Rock Material 30

Fissured Rocks 30

Discontinuities 40

Continuous and Discontinuous Rock Masses 47

Engineering Classifications of Rock Masses 49

3. PRII"CIPLES OF STEREOGRAPHIC PROJECTI ON AND JOINT SU RV EYS 58

Conformal Stereographic Projection 58

Basic Constructions 66

Joint Surveys and Statistics on the Sphere 83

Bias in Measurement of Joint Orientations from Drill Holes and Outcrops 86

The Directionality of a Jointed Rock Mass 88

4. EXPLORATION OF ROCK CONDITIOI"S 91

Geological Maps and their Interpretation 91

Aerial Photo Interpretation 104 Terrestrial Photographs 112

Geophysical Methods 121

Drill Holes 127

Absolute Orientation of Structural Features in Drill Core 142

xi

xii

5. MECHANICAL PROPERTIES OF DISCONTINUITIES 158

Determination of Properties 158 Deformations in Joints 170

Peak Shear Strength 183

Influence of Joint Orientation 200

6. APPLICATIONS OF STEREOGRAPHIC PROJECTION IN MECHANICS OF DISCONTINUOUS ROCKS 269

Introduction 209

Kinematical Considerations 210 Operations with Vectors on the Stereonet 217 Application of the Stereographic Projection in Defining a Wedge

Formed by Intersecting Discontinuities 225 Analysis of Rotation 231 Analysis of Sliding of a Block on a Plane-the Friction Circle Concept 237 Estimate of the Displacements of a Block Under a Dynamic Impulse 244 Slip of Tetrahedral Wedges 247 Sliding of Tetrahedral Wedges With Only One Free Surface 255 Slides Composed of Two Blocks 261 The State of Stress in Rocks 270 Conclusion 275

7. PHYSICAL MODELS 277

IKinematic Models 277 Physically Scaled Models 285

8. THE FINITE ELEMENT METHOD 300

Introduction 300 The Method 301 Formulation of Element Stiffness Matrices and External Loads 307

The Constant Strain Triangle 308 Initial Stresses in the Rock 317 Constant Strain Joint Element 320 Assembly of the Structural Equations 330

Iterative Solution to Simulate Real Properties of Joints 333 Sources of External Load 349

Example Problems 354

Incremental Loading 367

Rigid Block Analysis 368

APPENDIX ONE-AN ILLUSTRATIVE FINITE ELEMENT PROGRAM 369

Purpose and Scope of the Computer Program 369

Program Structure 371

Input Instructions 375 Examples of Input and Output 378

Listing of Finite Element Program-"JETTY" 394

APPENDIX TWO-CONVERSION FACTORS 417

REFERENCES 419

INDEX-Subject 451, Author 467

Contents

Methods of Geological Engineering in Discontinuous Rocks

1 i troduction

This book discusse s methods an d proce du r es ava i lable to assess

the i n fl uence o f discontinuities on the behavior of r o cks in engineer-

ing app l ications. Most rock masses in the region of influence of

works s u ch as qua r r i e s , road cuts, foundatio ns, dam abutments, tunnels,

and un derg r ound chambers contain planar surfaces of potential or real

weakn e ss. The se weakness planes come in all lengths and spacings and

have vary i ng deg r ees of i n fluence on the overall mass properties. We

rare l y can a f ford to c l ose ou r eye s to thei r presence in attempting

to cal culate roc k perfo rmance.

Use o f t he phrase "di s c ontinuous rocks" in the title implies

th a t there are other rocks wh ich are truly continuous. This is not

s t r i c t ly correct for even the mightiest wall of granite has exfolia-

t i on s u rfaces and other widely spaced joints and faults in various

orien t at i ons . There are ma ny rock masses, however, in which the

discontinuities t hough present a r e not the we akest link in the list of

compon ent s whi ch col l e ctively give the rock its strength and other

physical at tribut es. In friable sandstones of Tertiary age, for

examp l e , the s a n d g rains may be so poorly bonded that failure through

the rock mate r ial itself is more likely than failure by sliding on

bedding pl an e s or joints . This may also be the case in shales which

t hough dis r up ted and loosened by anastamosing cracks due to slaking

near exposed s u r faces, will tend to fracture through the body of

materia l rathe r t han on structural ly controlled surfaces. The y ounger,

2 Introduction

Figure 1-1. A discontinuous rock mass. Columnar joints and flow band-ing in a basaltic flow-Iceland; (courtesy of Dr. Tor Brekke).

weaker rocks as a rule tend more closely to fit a "continuous model",

while hard rocks invariably are controlled in their failure modes by

their pattern of discontinuities. How else could a rock like quartzite

fail? It has a compressive strength some ten times that of mass con-

crete.

Soils are not continuous materials; they have grains and pores.

But they have been successfully analyzed using a continuum model --

heterogeneous, if necessary, but continuous. The discontinuous rocks

with which we are concerned here might at first thought be likened to

soils and treated using soil mechanics theory and techniques. In

fact, some early attempts were made in this direction. But there are

fundamental differences. The discontinuous rocks have essentially no

pore space, except that of the rock material itself (pores in the

rock material are analogous to pores in the grains of the soil). Thus

the discontinuous rock is locked together into a perfectly fitted

pattern. To create failure, pore space must be created and this

implies dilatancy, or bulking in the construction man's parlance.

Not only normal and shear forces act inside such rock masses, but

moments as well. Soil grains may be free to turn in place; rock

blocks are not.

It might seem a hopeless quest to rationalize the design

process when dealing with such a material as discontinuous rock.

Sometimes it is hopeless, and only previous experience, or trial and

Introduction

error, can be used. Other times, fortunately more f req ue nt l y as we

gain experience, the network of discontinuities can be accurately

described and mapped and its influence on the mass behavior can be

adequately evaluated . The elaboration of these methods i s the subject

of this book. First we must meas u r e the orientations of t h e v a rious

sets of planes which penetrate the rock in quest ion. This can be done

by geological observations on outcrops, by i n s pection of n a tural and

artificial cuttings, by study of aerial photographs , by me as u r ements

on drill cores and the walls of exploratory borings, and by geophysi-

cal traverses using a number of avail ab le techniques. Chapt e r four

reviews some of the relevant methods of exploration aimed at p rovi ding

a description of the orientation and spacing of the discontinui ty

network in a rock body. Chapter five then introduces t he mechani cal

properties of surfaces of d i scontinuit y and consi de rs their meas ure-

ment and numerical values. When we are dealing with s ingle, very

important weakness surfaces, whose orientation and position with

respect to a project are known with precision, it is possible to make

explicit analyses of the resulting stresses and deformations ; this

can be done by kinematics and statics, using stereog r aph i c pro j ecti on

to handle the three dimensional aspect of the problem, as di s cus sed

in chapter 6. In chapter 7, physical model methods are introduced.

An emphasis is placed on kinematical models which examine t he v a rious

possible modes of failure of a discontinuous rock mass in an engineer-

ing context. Analyses can also be performed by numer ica l me thods;

the finite element method is introduced in chapter 8 and a dig i tal

computer program, designed so that it can be read along wi t h the

theoretical discussion of chapter 8, is presented in the Appendix.

Because stereographic projection p r o cedures are used fre quently

throughout the book as a means of solving spatial problems, such as

orienting planes in drill core, measuring angles on terrestrial photo-

graphs, resolving stresses on planes of given orientation, a nd operat-

ing with vectors , a chapter has been addressed specifical ly to tech-

niques of stereographic projection (chapter 3). In dealing with

vector quantities, we must use the whole sph e r e so the s ub ject i s

treated somewhat differently than in works on structura l geology .

Chapter 2, on classification of rock, has been written to r elat e the Goodman-Geolog ical Eng. - 2

3

4 Introduction

d i scontinuous rocks to other categories of rock J i .e. to se t thi s wo rk

i n its proper context .

Geological ngine ering is concerned with a broa d spect rum of

natural proces ses. At one end of the spectrum are those geologic

hazards , such as large landslides, active fau l t s, an d cavernous

terrain , which dwarf an intended project in terms of size, potential

energy , or the cos t of neutralizing the haz ard; with such h az ards,

the geological eng i neer can do l ittle more than recog ni ze, describe

and be responsive to eventualities. He uses vari ous methods to

study thei r potential and t o observe the i r acti v i t y , but h e has little

effect on t he phenomena t hemselves. At t h e o ther en d of the spectrum

of geological engineering applications are min i ng a nd quarrying

activities where the geology i not only studie d an d eva l uate d , but

Figure 1-2. A concrete arch dam. Mossyrock Dam, Cowlitz River, Washington: a doubly curved, thin arch dam 365 feet above riverbed, 606 feet above the basalt bedrock; (courtesy of Tacoma City Light).

Introduction

Figure 1-3. Malpasset Dam site, looking at the left abutment an d into the reservoir area.

wherein the r ock is removed, crushe d , st o ckpiled and perhaps even

emplaced i n a hosti l e and caus ti c environment, for examp le , as

aggregate in cement. In between are those constructions and excava-

tions, such as dams, underground openings and open cuts, which apply

static or dynamic loads or un loads at the surface or subsurface. It

is with these that the methods discussed in this book are primarily

concerned.

Large dams, especially con c rete ar ch dams as in figure 1-2

combine large loads wi t h the hydraulic and chemical effects of water

and therefore place challenging demands on geological engineering

investigations. Much of the recent interest in geological engineering

and rock mechanics has in fact been motivated by concern about the

safety of dams and reservoirs following the catastrophes at Malpasse t

dam in France and Vajont reservoir in Italy. At Malpasset dam,

5

fi gure 1 - 3, a complicated set of circumstances deriving from the

behavior of the schistose gneiss bedrock caused a rupture of an ar ch

dam. French investigators determined that a wedge of rock in the

abutment, bounded by intersecting weakness surfaces, moved due to the

thrust of the dam and high water pressure within the abutment (Bernaix,

196 6) *. The high wate r pressure was generated by the development of

* References will be fo un d in the Bibliography, on page 419

6 Introduction

Figure 1-4. Kukuan Dam, during construction; (courtesy of Coyne and Bellier).

a natural flow barrier under the line of action of the dam as fissures

within the rock mass closed in response to applied load. The Vajont

failure (Muller, 1964 and 1968) occurred when a massive landslide

moved on bedding surfaces into a relatively small reservoir, causing

overtopping and flooding. The landslide was triggered by uplift

forces associated with reservoir filling.

The large influence of discontinuities on construction operations

in rock is well illustrated by the Kukuan arch dam, designed by

Coyne and Bellier for Taiwan Power Company. This dam, 86 meters high,

was constructed in a valley cut 500 meters deep into alternating

layers of slate and quartzite. Thin clay seams containing graphite

compromise the stability of unfavorably oriented layers of the site.

The right bank (figure 1-4) is a 60 - 70 degree dip slope. To

found the dam in solid rock, it was necessary to excavate through 20

to 40 meters of loosened slabs, but conventional excavation was

undesirable because of the slide potential. Grouting and "dental

work" (localized replacement of weak rock with concrete) were un-

successful. A solution was obtained in which tunnels up to 10.7

Introduction

meters wide were driven well into the abutments and backfilled with

concrete. Since the tunnels cut across the bedding, they were stable.

After driving a tunnel to the full depth and width, it was concreted

to within several meters of the crown. Then, after two to three

weeks, a stone protection was laid on the concrete fill and a t unnel

was excavated above. The process was repeated until eight tunnels

had been constructed, producing a stable concrete structural abutment.

Activities in advancing the construction and utilization of

tunnels and underground chambers have also created interest in methods

of geological engineering. Investigations of tunnel sites remain

fairly primitive because the sites are long, and remain inaccessible

until construction. Some attention has been focused on assess ing the

excavatability of the rock from tests on samples, but geological and

geophysical prediction techniques, and analytical methods to forecast

formation conditions are not yet generally available.

7

Techniques for investigating and anaJyzing rock behavior for

underground works such as subsurface power plant chambers (figure 1- 5) ,

(b)

(a)

Figure '5. (a) Oroville Dam project. The dam has sh ells of gravel while the core is derived from a vast alluvial fan; (courtesy Calif. Dept. of Water Resources). (b) Oroville underground power station machine hall during construction. The man standing in the lower left gives the scale; (courtesy Calif. Dept. of Water Resources).

8

Figure '6. Spillway excavation on left abutment of Chivor Dam, Colombia. Notice the truck and shovel for scale. The smooth surface of discontinuity in the middle left was exposed during construction and caused a design change. The benches are 5 meters wide and spaced every 10 meters; (courtesy of I ngetec Ltda., Bogota).

Introduction

subterranean factories, defense installations, storage chambers, and

mine shafts, on the other hand, are better developed. It is usually

feasible in such projects to make detailed investigations including

determination of rock properties, analysis, and instrumentation. An

additional aspect of investigations for underground structures not

addressed in investigations for dams, is the role of in-situ stresses.

At great depth, such as in some mines in South Africa and Canada, one

occasionally reaches the natural strength of the rock.

Surface excavations for spillways (figure 1-6), mine pits

(figure 1-7), transportation routes, power plants, and for access to

the underground, are other important areas of rock engineering. In

mines, important savings in excavation volumes can be achieved by

application of simple theory supported by field observations of geo-

logical details, back calculations of failures, piezometric measure-

ments, and analysis of the response of instruments (Hoek and Bray,

1974). Careful blasting practise and instrumentation can insure safe

operation of engineering works immediately adjacent to rock slopes,

which themselves can be regarded as engineering structures (figure 1-8).

Though the specific choices of methodology will differ among all

these types of projects, basic similarities of purpose prevail.

Introduction

Figure 1-7. Chamblshi Mine, Zambia (courtesy R.S.T . Ltd. and Prot E. Hoek) ,

First .. the geo logy of the site must be defined; this entails mapping

of f i e l d exposures J study of aerial photographs, and specific explora-

t ion with excavations o r drill holes . Then . the properties of the

rocks must be assessed. Here there can be different choices of

Figure 1-8. Pre-split rock exca vation for Stockton Dam; (courtesy Mr. Lloyd Underwood, Corps of Engineers, Missouri River Division).

9

10 Introduction

methods since the relevant properties to be evaluated vary greatly

according to the purpose of the project. The behavior of a complex

of underground openings reflects the initial state of stress; in

some analytically based design processes the in-situ stress will need

to be measured, or otherwise determined. In shallow rock excavations,

on the other hand, the shear strength and water pressure levels are

more critical, while for foundations the deformability of the rock is

foremost. Thirdly, through model studies, computer analysis, or

reference to appropriate similar experiences, the response of the work

at the specific site with the assigned properties is evaluated. If

unsatisfactory, the structure may be relocated or the properties may

be changed in some measure by excavation, grouting, drainage, bolting,

or other means. In this case new explorations, tests, and studies

will be inaugurated. The designer will have the most economical

solution if he is able to adapt the style of structure to the par-

ticular attributes of the site, most of which have been provided

naturally. The methods and work of geological engineering are there-

fore mainly devoted to discerning just what is already there.

The nature of rock is vastly different from other types of

engineering materials. Therefore it is natural that the methodology

employed for its characterization should be peculiar to the field of

geological engineering. Nevertheless, each of the methods employed

and discussed here has its cousins in other disciplines, and a book

such as this must cross the borders of many fields. These include

mining; petroleum; geophysics; cartography; planning; soil mechanics;

hydraulics; mechanics of materials; concrete technology; structural

engineering; statistics; aeronautics; and computer science. The

obvious consequence is that sources of literature of interest for

further reading are scattered among numerous journals, and reference

books. However, a number of basic references and journals can be

singled out as especially relevant. These are listed in Table 1-1.

Introduction

TABLE 1-1

Some Sources of Information

Bibliographies and Abstracts

KWIC Index of Rock Mechanics literat u re publ ished befo re 1969 -2 volumes. Produced by Rock Mechan ics In formation Se rvice, Imperial College, London. Published by AIME, 345 East 47th St., New York, N. Y. 10017

Geomechanics Abstracts - Part II of the Inter. J our. Rock Mechanics and Mining Science Published by Pergamon Press from volume 4 ( 1973 ) o nwa rd (Originally called Rock Me cban i c s Abstracts; pro duced by Imperial College).

Geotechnical Abstracts - Monthly with ann ua l indexes Deutsche Gesellschaft fur Erd- und Grun db a u ( fo r I nt e r. So c. for Soil Mechanics and Foundation Engine e r ing). (Published also in a card format called "Geo dex Retr i e val System " ) .

Bibliography and Index of Geology - Monthly Geological Society of Ameri ca .

National Technical Info rmation Service, Sp ringfie ld , Va . 22 151 (Bibliography and source for U.S. Government documents).

Geoscience Abstracts - Monthly . American Geological I nsti tute, Washington 25 D.C. (A special supplement is devoted to a "Bib liograph y of bibliographies of the States").

Chronique des Mines et de la Recherche Min ie re published 10 times per year by Centre d'et u des geolog iques et Minieres

Annotated Bibliography of Economic Geology - semi - annu a l. Economic Geology Publish i ng Co.

Journa ls and Serials

Rock Me chanics ( I n ter. Soc . for Rock Mechanics) Forme r ly "Rock Mechanics and Engineering Geology" .

International Journal of Ro ck Me chanics and Min i ng Science (Pergamon Press).

Eng ineer i ng Geo logy (Elsevier).

Quarterly Journal o f Engineering Geology ( Geolog i c a l Soc. of London) .

11

12

Bulletin of the Assc~iation of Engineering Geologi s ts.

U.S. Bureau of Mines , Reports of Investigations and other publications.

Canadian Geotechnical Journal .

Geotechnique .

Bulletin of the Inter . Association o f Engineering Geo logy.

In troduction

Proceedings of Congresses and Symposia of the International Society for Rock Mechanics*

First Congress - Lisbon 1966 - 3 vo l umes.

Second Congress - Belgrade 1970 - 4 volumes

Third Congress - Denver 1974 - 5 volumes.

Symposium on Rock Mechanics - Madrid 1968 - 1 volume.

Symposium on Stress Measurement - Lisbon 1970 - 1 vo l ume .

Symposium on Large Permanent Underground Openings - Os lo 1969 -1 volume

Symposium on Rock Fractures - Nancy 1971.

Symposium on Percolation through Fractured Rock - Stuttga rt 1972.

Proceedings of Symposia on Rock Mechanics-U.S.A.

8th to 12th, 1966 - 1970 (AIME).

13th to 15th, 1971 - 1973 (ASeE).

Previous Symposia are l isted in preface mater i a l f or above Symposia .

Other proceedings of interest are lis t ed in volume 2 o f KWIC Index, (see "Bibliographies and Abstracts I I abo ve) .

Textbooks

Coates, D.F., (1967) "Rock Mechanics Principles" , Canadi an Dept . Energy, Mines and Resources, Monograph 874.

Hoek, E., and Bray 1 J. (1974) ItRock Slope Engineering'!, (In s t. of Min and Metal, London) .

*Can be ordered through ISRM , Laboratorio Naciona1 de Engenhari a Civil, Avenida de Brazil, Lisbon, Portugal.

Introduction

Jaeger, J.C., and Cook, N.G. W. (1969) "Fundamentals of Rock Mechanics", (Methuen).

Krynine, D., and Judd, W. (1959) "Principles of Engineering Geology and Geotechnics", (McGraw Hill).

Ob e rt, L., and Duvall, W., (1967) "Rock Mechanics and the Design of Structures in Rock", (Wi l ey).

Scott, R.F., (1963) "Principles of Soil Me ch anics", (Addi s on Wesley) .

13

2 rock classification

While this book primarily concerns the discontinuous rocks, it

is necessary to see this rock class in context and accordingly the

question of rock classification in general will be explored. The

object of rock investigations and measurements is to make judgments

about the rock as a prelude to some action. The properties used to

classify the rock will vary according to the designer's purposes and

may include various subsets of: shear strength; flexural strength;

tensile strength; elasticity; permanent deformability; creep-rate;

water flow and water storage properties; in-situ stress; drillability;

fragmentation characteristics; and sometimes density, thermal expan-

sion, mineralogy, and color.

THE NATURE OF ROCK

One can not assign rock properties to a design calculation with

the same degree of certainty as with some other types of engineering

materials. The reason is that there is rarely a wholly dependable

large sample of the total population available from which test results

can be extracted. The application of principles of structural geology

makes the sampling problem solvable. But we must realize that most

of the volume of rock of immediate concern is hidden and inaccessible

and, unfortunately, what we do see is rarely representative of what we

don't. It is almost a law of geological engineering that the hidden,

Rock Classification 15

mantled material is the weakest and potentially most troublesome ; only

the sandstone layers will cropout in a formation composed of sandstone

and shale; only the flow rocks will form ledges in a volcanic series

of basalts and pyroclastics. The granite will form a hill, but the

fault zone through it will form a valley.

Nor can the designer of a work in rock make use of rock proper-

ties with the same rigor as he might for other types of st ructural

and hydraulic computations, because rocks seldom lend themselves to

the usual sort of idealizing assumptions. First, most rock f o rmations

have directionality, such as bedding in sedimentary rocks, flow band-

ing in volcanic rocks, and foliation in metamorphic rocks, and are

consequently moderately to highly anisotropic. Then we fin d rock

responding differently to excavation according to the i n it ial state o f

stress , particularly in underground applications, and this is heavily

dependent on the stress history which will be known only occasionally.

Many rocks are semi -discontinuous on the hand specimen scale owing to

a network of fissures and flaws, and almost all rocks on the formation

scale are penetrated by surfaces of potential or real discon t inuity.

At the depths reachable in deep mines, deep drill holes, and some

tunnels, some rocks are ductile, and very few rock s behave en t irely

elastically even at low pressures. Some rocks are chemical l y change-

able within the lifetime of an engineering work a n d even more show

great variability vertically and horizontally, due to different deg rees

of weathering. In the face of these difficulties, results of comp u t a-

tions are to be utilized with restraint, and controlled by observ a-

tions during construction. Fortunately, it is often sufficient for

engineering purposes to produce only a reasonable estimate of the

final behavior - an estimate that can be arr i ved at sati s f actor i ly by

rock classification.

ROCK SPECIMEN VERSUS ROCK-MASS

In a discussion of rock classification, we must carefully dis-

tinguish characteristics of a specimen of rock from properties of a

body of rock in situ which, in the language of rock mechanics, we c a ll

the rock mass. The mass is comprised of the rock, its network of

discontinuities and its we a t hering profile. The behavior of the rock

16 Rock Classification

mass reflects all of these components as well as water and stress

regimes, strength, deformability, and permeability, which may be

largely unrelated to material properties.

Classification of the entire realm of rock masses for the to-

tality of applications would demand an unwieldy number of independent

factors because different pursuits require different parameters. In

assessing the suitability of facing stone, aggregates, embankment

materials, and other rock products, we need rock specimen attributes

describing durability, strength, thermal expansion, shrinkage, swell,

absorption, and specific gravity. Rock mass characteristics affect

items related to the cost of production. In regard to excavations,

both specimen and mass characteristics are essential, the former

affecting drillability and durability and the latter being basic to

stability while also influencing excavatability. The essential

factors for foundations, particularly for hydraulic structures, are

those descriptive of rock mass deformability, stability, and per-

meability which derive principally from the discontinuities, (although

rock specimen .characteristics may sometimes control the design, as

for example in non-durable, fissured, weathered, or permeable rocks).

First, we will examine classification of rock specimens then

the weathering profiles and systems of discontinuities and finally

the classification problem for rock masses.

PETROLOGIC CLASSIFICATION OF ROCK SPECIMENS

Geological methods of classifying rock specimens are based on

a number of different criteria, which can be studied in Williams,

Turner and Gilbert (1958). We will explore the wisdom of using geo-

logical rock names and petrological descriptions for engineering

purposes.

A description of a rock's texture and fabric affords a basis for

understanding its mechanical properties, which are closely related to

interparticle bonding, interlocking and imperfections. The crystalline

rocks (figure 2-1a) have tightly interlocked particle arrangements

sometimes impaired by micro fissures within and between crystals.

Coarse grained crystalline rocks tend to be weaker and less stiff

than fine grained or aphanitic crystalline rocks. Foliation, the

Rock ClassificatIOn 17

(a)

(b)

(c)

Figure 2 1. (a) High ly inter locked, crysta ll ine texture of a Mesozoic quartzite; (courtesy f Prof R. WenkL 25.5X . (b) Highly anisotropic crystal l ine texture with oriented fissures (fracture 1.ledvage) III chlorite schist- Homestake M ine, S.D.; (courtesy of Dr W Chinn). 40X (c) Porous, clastic texture ' eolian sand, stone from Olduvai gorge, Tanzan ia, consisting of poor ly sorted rock fragments and grains, some coated with clay. Cavities occur in altered nepheline grains (N); (courtesy of Prof. R. Hay). 136K


most predominant fabric element of metamorphic rocks, causes strong

anisotropy and surfaces of weakness within the scale of the specimen

(figure 2-1b). Foliation is particularly pronounced when formed by

coplanar platy minerals like mica. In clastic rocks (figure 2-1c),

grain size has far less influence on mechanical properties than the

nature, strength and durability of the binder or cement. Properties

of cemented and compacted varieties of shale, for example, can be as

different mechanically as soil and rock. Bedding is the most im-

portant structural feature of sedimentary rock on the specimen scale,

as well as in the rock mass; it creates anisotropy in all properties.

Since geological names for rocks are intended to classify rocks

according to differing modes of origin, one may wonder if they are

meaningful for geological engineering practice. In the igneous rock

group, the genetic division between intrusive and extrusive rocks is

meaningful in terms of engineering attributes since it concerns the

depth of formation. Features derived from the surface environment --

vugs, amygdules, and flow structures -- partly determine the mechani-

cal properties of the volcanic flow rocks. The plutonic rocks, on

the other hand, present quite different aspects linked to their

formation at depths of perhaps 30 miles where the pressure approxi-

mates 150,000 psi (1000 MN/m2 ). For example, plutonic rocks such as

granite may possess large horizontal stress and fissures from un-

loading and a strong inclination towards chemical weathering.

Dynamically metamorphosed rocks (as opposed to products of thermal

metamorphism) contain miniature fold and fault structures and minerals

oriented during growth under deviatoric stress. The various genetic

processes responsible for the sedimentary rocks also produce distinct

assemblages of properties linked to the mode of origin, -- although

on the specimen scale the mechanical properties are more directly

related to textural and mineralogic considerations independent of

origin.

Mineralogic classifications form the basis for the actual rock

names in the igneous and metamorphic rocks and to some extent in the

sedimentary rocks. The mineral composition of crystalline rocks is

not vital to a classification of mechanical properties and consequent-

ly many of the rock distinctions important to petrologists are useless

Rock Classification

for engineering purposes; for example we usually don't care whether

a rock is classified as a granodiorite or diorite or tonalite.

However, the accessory minerals may vary from one species to another

and these, more than the proportions of quartz and feldspar may

affect engineering response. Pyrrhotite (Martna, 1970), possibly

pyrite, iron-rich micas, nepheline, leucite and nontronite have been

identified as instrumental in deterioration of originally solid rocks

quarried for aggregate and building stone. Minerals containing vugs

filled with carbon dioxide can lower the pH of the groundwater, con-

tributing to rapid weathering as at Bergeforsen Dam (Aastrop and

Sallstrom, 1964). Any of the sheet silicate minerals, e.g. mica,

chlorite, talc, and serpentine, introduce low shear strength,

especially if in coplanar orientation; mica schist, serpentine and

talc schist can be hazardous rocks in foundations and excavations.

Glass, and secondary minerals zeolite and opal, can promote chemical

reactions with cement even when present in small quantities in rock

aggregate. Crystalline sedimentary rocks include some varieties

largely or partially composed of weak, soluble, or non-durable grains,

e.g. clays of the montmorillonite group, gypsum, halite, sylvite,

we ak shales, coal, chalk and chert.

In summary, though the science of petrology has evolved accord-

ing to the needs of classical geology, its refined terminology and

class distinctions are frequently meaningful for engineering work.

Moreover, as geologists are familiar with it, and its rock classes

are generally mappable, the geological nomenclature, especially when

accompanied by textural descriptions and mineralogic details, is

the most appropriate rock material classification for engineering

purposes. The complete classification of the rock material must of

course describe the state of weathering, the durability, and the

degree of fissuring.

ROCK VERSUS SOIL AN D WEATHERED ROCK

The most vital distinction to be recognized is between rock,

weathered rock and inherently soil-like rock. The distinction is

essential for all engineering work in rock, and yet it is not an Goodman-Geo log ical Eng.-3

19


elementary proposition. The fundamental precept is that to be rock,

the material must be strong and durable. It is solid when first

encountered and can not be softened, disaggregated, or' easily weak-

ened by accelerated weathering. Furthermore, it does not swell or

shrink appreciably upon soaking. These requirements are pragmatic but

do not coincide with geological nomenclature, in which a rock is

defined as "any consolidated or coherent and relati vely hard,

naturally formed mass of mineral matter".*

Table 2-1 was based upon one devised by Karl Terzaghi for

students in his engineering geology class** to distinguish between

rock, weathered rock, and soil-like rocks. One may apply the term

"solid rock", according to Terzaghi only if a rock is solid with

a ringing sound when struck by a hammer and remains solid throughout

weathering tests and soaking. Moderately soluble varieties, such as

limestone and dolomite, will still be classified as solid rock, but

greatly soluble rocks such as salt and gypsum will not survive a

reasonable weathering test intact. Rocks which are originally solid

but break up into small, hard pieces with a clean surface on weather-

ing are termed fissured or crushed unaltered rocks, whereas if the

rock disaggregates or yields greasy surfaces, it is an unstable or

slightly decomposed rock. If such a rock exhibits perceptible

volume change upon soaking, Terzaghi thought "rock" would be a

dangerous misnomer; he preferred to designate swelling materials as

"intermediate between rock and clay, rock characteristics predominat-

* Dictionary of Geological Terms" Dolphin Reference Book C36D. The above is the ordinary usage but this dictionary gives as a strict definition "any naturally formed aggregate or mass of mineral matter whether or not coherent, constituting an essential and appreciable part of the earth's crust". The word consolidated in the first definition is troublesome to engineers familiar with soil consolidation theory which refers to the expulsion of water from the voids of a soil under pressure. The geological usage means firm.

** Table 2-1 is based upon one devised by Karl Terzaghi and distributed to students in his course on engineering geology at Harvard University in the 1950's. A copy revised shortly before his death was generously supplied by Dr. Ruth Terzaghi. A somewhat similar approach is used by the National Institute for Road Research, South Africa, as published by Weinert (1964); see Fookes, Dearman, and Franklin (1971).

TABLE 2-1 EF FECTS OF SATURATION ON ROCKS AND ROCK-LIKE MATERIALS

Terzaghi's Guides for Distinguishing Rock, Weathered Rock, and Soil *

In original

state

Solid wit h ringing sound when struck with a hammer

Solid with dul l sound when struck with a hammer

Afte r r e peated drying, i mme r sing, and sha king , or upon prolonged exposure to the atmosphe re

unchanged

bre aks up i nt o small h ar d pieces with clean surf a ces

break s up into small fragments with "greasy " s urfa ces owi ng to the

Vo l ume change produced by s a t ura ting dri ed f ragments wi t h water

presence of fi ne - imperce p t i b l e graine d weathe r i n g product s

breaks up i n to indi-v idual sand or s ilt part icles

break s up i n to small angular fr agments with-out a ny indi cation of che mical a lterat ion

gradual ly t ransformed into a suspension of soi l p art icles

gradua l l y t ransformed i nto a s us pens i on of c l a y part i c le s a n d a sedime n t consis ti ng of angular rock f ragments

comp l etely t ransforme d into a s usp e nsion and / o r a loos e sedi ment

measurable

impercept ibl e to importan t

Group

a) so l i d rock

b) f ine ly fi ssured or crushe d unal tered r ock

c ) s lightly de -compose d f issure d rock

d) sandst one or muds t on e wit h un s t ab le ce me nt

e) i nterme d iat e between r ock a nd c l ay, rock ch arac t eristics domi nan t

f) i nterme di ate be twee n rock a n d c l ay , clay c harac teristics dominant

g ) thoroughly decompose d r o ck

h ) c l a y , silt , a nd very fin e san d in dr y o r a very com pacted con di tion

* From Professor Kar l Terzaghi' s 00urs e notes fo r Enginee ring Geology a t Harva rd Uni versi ty; i n clude d wi t h k ind p ermission o f Dr . Ru th Terzagh i (w i th minor e di t orial c h ange s) a nd including revisions made by Kar l Terzaghi shortly b e f ore his death.


ing". Materials that are not solid with a ringing sound when struck

by a hammer when first encountered should not be referred to as rock

at all, according to this scheme. Many sedimentary rocks would

accordingly be termed "soil-like rocks II in maps and reports, and the

resulting impression would be correct for the engineer.

Geological investigations must correctly diagnose a specific

soil-like condition as either inherent or localized. Weathering, and

hydrothermal alteration -- the first usually intensifying towards the

surface and the other, with depth or laterally -- may produce spotty

and variable degrees of localized softness. In contrast, some

sedimentary rocks are inherently soft either through incomplete

cementation, intense fissuring, or regional alterations; neither

"dental work" nor outright "extraction" can improve the rock condi-

tions in this case.

The distinction between rock and soil is especially important as

regards specifications for excavation contracts. So many legal

controversies have revolved about this point that agencies such as

the U.S. Bureau of Reclamation have been forced to adopt almost

comically detailed wording for contracts, as in Table 2-2. The main

ideas are that the material to be excavated is rock only if it is

both in place, (or of large mass) and solid. If it is too risky to

attempt a classification, the excavation receives one name--unclassified

excavation -- and one price throughout. This can happen in deeply

weathered materials, with their extreme variability and gradational

qualities, in soil-like soft rocks, in bedded rocks alternating in

hardness, and in very dense or cemented soils.

WEATHERING

Closely related to the question of differentiating soil from

rock, is evaluation of the degree of weathering of the rock material.

The importance of the subject is suggested by a voluminous literature,

a selection of which is included in the list of references. Rocks

respond to prolonged weathering in many ways. The granitic rocks

become cracked and then decayed by the carbonic acid developed as

rain water filters through the soil; this reagent attacks the feld-

spars and dark minerals releasing soluble salts of K, Mg, Fe, and Na,

Rock Classification

TABLE 2-2

Classification of Excavation According to U.S. Bureau of Reclamation Contract Specifications

"Except as o therwise provided in these specificat ions, ma te r ial excavated will be measured and classified in e xcavation, to the lines shown on the drawings or as provided in these specificat ions, an d will be classified for payment as follows:

Rock Excavation. For purposes of classification of excavat ion, rock is defined as sound and solid masses, layers, or l e dges of mineral matter in place and of such har dn ess and t extur e that it:

(1) Cannot be effectively loosene d or broken down b y ripping in a single pass with a late model trac t or-mounted hydraulic ripper equipped with one digg i ng point of standard manufacturer's design adequately s i z ed for use with and propelled by a crawler-type tract or rate d between 210- and 240-net flywheel horsepower, ope rat i ng in low gear, or

(2) In areas where it is imp r acticab le to class i f y by use of the ripper described above, rock e x cava tion is defined as sound materi a l of such hardness a n d t e x t ure that it cannot be loosened or b roken down by a 6 - pound drifting pick. The drift~ng pi ck s hall be Class D, Federal Specification GGG-H-506d, with handle not less than 34 inches in length.

All boulders or detached pieces of solid rock more th a n 1 c ub ic yard in volume will be classified as rock e x cavat ion.

Common Excavation. Common excava~ion i nc ludes a ll materi a l other than rock excavation. All boulders or de tac he d piece s of solid rock less than 1 cubic yard in vol ume will b e class i-fied as common excavation."

as well as free silica which may be transported out of t h e weat her i ng

environment , and detrital clay and resistant q uartz grain s whi ch

usually remain. The rock is gradually transfor me d into a "saprolyte fl ;

figure 2-2a, which resembles rock but has the st reng t h o f a dense

soil. Vargas (1953), Ruxton and Berry (195 7 ), Lumb ( 196 2) , Dee r e

and Patton (1971), and others have described the transi t ional states

23


(a) (b)

Figure 2-2. (a) Decomposed granite. Former joint blocks contain hard " core stones" in their centers while relict joints are now sandy clay seams or partings in the thor-oughly weathered rock (saprolyte). (b) The top-of-rock surface in soluble marble; Columbia, California. The soil was removed by hydraulic monitors to obtain placer gold.

of granitic rocks and their properties. Basic igneous rocks follow

a similar sequence but tend to produce a residuum richer in clay. The

soluble rocks become enriched in impurities, often clayey, and develop

stable or unstable vugs according to their strength. Limestones are

often karstic whereas gypsum and halite are simply thinned or removed

altogether, the karstification inducing almost immediate collapse,

(Brune, 1965), (figure 2-2b). Intermediate weathering states con-

sisting of vuggy rock are less common in gypsum than in limestone

and dolomite rock. Anhydrite expands, relative to the initial solid

volume, as it is converted to gypsum by hydration (but relative to

the total volume of reactants it contracts).* Compacted shales and

* There is a difference of opinion on hazards of anhydrite. A thesis by Sahores (1962) considered the engineering problems implied by volume expansion to be overstated. Brune (1965) on the other hand, described uplifts and explosions occurring naturally in an area of West Texas underlain by anhydrite at depth; moreover the anhydrite grades into a thinner, folded gypsum layer updip and the uplifts occur directly over the locus of anhydrite - gypsum interfingering leaving no doubt that con-version of anhydrite to gypsum is responsible for these violent phenomena.


poo rly - cemented sandst ones -- t he soil-l ike r ock s -- disaggregate

an d return to sedimen t i n re s ponse to weathering, and montmorill on itic

v arieties swel l . In genera l , as t h e non-soluble rocks pass through

intermediate stages of we at hering they gain porosity and deformability,

l ose strength a nd e las t ici t y , a n d become first more and then less

perme able; (see f or example Iliev's (1966 ) di scussion of property

cban ges in weathered monzon i te ), To classify t he materials and

a tt rib utes of t he wea t hered z ones, one must consider t wo i n dependent

c r it eria: fi r st it is ne cessary t o distingui s h differing degrees of

wea t hering of t he r ock itself; t he n th i s d i stinct ion must be super-

i mpose d on a class ifi cation of dif f ering s t yles and arrangements of

t he weathe rin g produc ts.

The Degree of Weathering

App r aisal of the deg ree of we a ther i n g actua l l y attained b y a

p a rt icular specimen of roc k mater ia l is bas ic to any meaningful

c lassifi cat ion of roc k masses with in t h e weat hered zone, which in the

trop ics a nd in par t icularly suscept ible rocks s uch as granite, may

exten d more t han 300 feet b elow t he g roun d surf a ce. No single index

de rive d f r om Simpl e field ob servat i on s o r labo r atory tests can be

e xpe c ted t o a pp ly a pp r opriately f o r all ma ter ia ls in the vast range

of weathering products de ri vable from intermediate stages of decom-

pos i t i o n o f rock , Several app roaches useful in particular rock types

a r e offered as e xamp les to b e emulated i n principle or detail as the

cas e warrants.

Lumb (196 2 , 1965) discussed correlation between soil and rock

p rop e r ties in g ran ites of Hon g Kong ordered by a mineralogic weather-

in g inde x Xd , Lumb's inde x , appropri ate f o r quartz b e a ring granitic

rocks i n which the felds p a rs are attacke d du r ing the decomposition

p roce ss , is assesse d by han d lens exami nati on of weathered and fresh

rock to dete r mine the percentages of felds par and quar t z as follows :

eN - N ) / (1 - N ) q qo qo

N is the weigh t r a tio of quar t z to quartz + feldsp ar in the weathered q specimen , an d Nqo i s t he corre spondi n g r atio i n th e intact , un-

weathered s pecimen . N is of th e o rder of 1/3 for a fresh granite qo and increases t owar d 1 as t h e weathering progresses. Thus the index


varies over the range 0 to 1 with an increasing degree of weathering.

Ege (1968) used a similar approach as one of four classification

indices for granitic rocks at the Nevada Test Site. The degree of

weathering is expressed by estimating the percent of altered minerals

in the rock, without reference to an unweathered standard. The

rock is classed as unweathered, slightly weathered, moderately

weathered, or severely weathered respectively as the percent of

altered minerals falls within the classes 0-10%, 10-25%, 25-75%, and

75-100%. The degree of weathering can also be classified on less

formal divisions as in the example by Kiersch and Treasher (1955) for

granodiorite at Folsom dam, California where: fresh rock was totally

unaltered; slightly weathered rock showed slight fissuring in the

feldspars and bleaching of their original color; moderately weathered

rock showed more intense bleaching and fissuring in the feldspars,

bleaching of the biotite, limonite appearing as specks and coatings

of other minerals and slight rounding of quartz grains; and highly

weathered rock showed strongly bleached biotite, the feldspars highly

fractured and bleached, the quartz grains highly rounded, and

limonite common as an accessory; further, the highly weathered rock

could be scratched readily with a steel nail. This simple classifi-

cation could be mapped and was successfully correlated with variations

in resitivities, seismic velocities, drilling rates with diamond and

percussion drilling, blasting patterns and powder factors, rippability,

grout takes, and suitability of stone for rock fill and rip-rap.

Iliev (1966) introduced an index K based upon the reduction of

longitudinal wave velocity with weathering.*

K v - V

o w V o

(1 )

The subscripts 0 and w identify the unweathered and weathered states.

Like Lumb's index, this one goes from 0 to 1 as weathering progresses.

* Such an index can be applied in the laboratory or in the field; in the latter case characteristics other than specimen properties are involved and classification by the application of this simple parameter can be wrong.

Rock Classification

Hamrol (1961) proposed a simple measurement of apparent porosity

by the water content of a rock (dry weight basis) as an index of

degree of weathering after quick immersion. The water content is

determined after oven drying at 105 degrees centigrade. Lumb (1962),

Pender (1971) and others have shown that porosity increases with

weathering (see figure 4-16); since some engineering properties of

rocks are directly associated with porosity or indirectly sensitive

to its changes (Griffith, 1937), it is not surprising that Hamrol's

index has met with success (Serafim, 1964) in recognizing rock grade

boundaries within a single rock type at a single engineering site and

in extrapolating results of field tests from one part of a foundation

to another. There has been little quantitative work on the changes

in properties of joints resulting from weathering.

The Profile of Weathering

Most engineering projects involve rock work in various levels

within the weathered zone, which may extend as deeply as 100 meters

below the surface. The outstanding feature of the weathering zone is

extreme variability of rock quality, both laterally and vertically,

(figure 2-2a); rocks of various degrees of weathering grade into one

another insensibly. Classification of the weathered rock mass can

27

be meaningful if described in terms of percentages of various weather-

ing products at any given level (weathering horizon). Deere and

Patton (1971) reviewed the weathering profiles of different rock

types and suggested standard terminology based upon the approach used

by Ruxton and Berry (1957) for granite soils of Hong Kong. These

papers, as well as the work of Fookes and Horswill (1970) Spears and

Taylor (1972), and others listed in the references should be consulted.

Durability

The discussion of weathered rock has considered only observed or

measured attributes of a present sample. What will the properties

be some years later, in response to construction and service? The

question of durability and its inverse, weatherability, is only

beginning to be answered by testing techniques and comparative data

meager data in view of the variety of engineering requirements. Some


of the minerals suspected of contributing to weatherability in rocks

were listed earlier; now, we will consider a simple index test.

Franklin's slake durability test. Fookes, Dearman and Franklin

(1971), and Franklin (1970) developed a durability test consisting

of a standardized measurement of the weight loss of rock lumps when

repeatedly rotated through a water-air interface. Ten lumps of 40

to 60 grams each are oven dried and weighed, and then placed in a

standard test drum (figure 2-3a) whose circumferential wall is

constructed of sieve mesh (2 mm opening). The drum is rotated at

(a)

(b)

Figure 2-3. (a) Slake durabil ity appara-tus (courtesy Soil Mechanics Equipment Co., Glen Ellyn, Illinois). (b) Franklin Point Load Testing Device (courtesy Soil Mechanics Equipment Co., Glen Ellyn, Illinois).

Rock Classi f ication

20 revolutions per minute for ten mi n u tes. The s l ow speed r e duces

me chanical wear effect in the agi t ating process. The dry we igh t

reta in ed a f ter t he we atherin g cycle, exp re s s ed as a p e rcen t age o f

the original wei ght, i s r epo rted as t he S l ake Durabil it y I ndex (I d ) .

Gamb l e (1 9 71 ), who evalua te d t h is i ndex in r el a tion to ot her dur-

ab i Ii ty and abrasi on tests used for aggregates , f ound t he sl ake

durabi l ity to be f ar gentle r a n d bet te r able to c ope wi th t he large

range of durab i li ty respon se offere d by ro cks (he pre f erre d a

modified durabili ty i nde x b ased on 2 c y c les of rot ation and drying ) .

Tests such as the Los An geles ab rasion tes t * a re mor e sensiti ve to

slight variations in durab i lity among ro cks to be considered for

aggregate.

To assess t h e weatherabi l i t y o f a ro ck , it i s me an ingful to

attempt t o simulate a project' s anti cipated we atherin g e n v ironment

at an accelerated rate. The pitfall is that un less one is ab l e to

incorporate all pert ine n t factors i n t he l aborato ry s i mulat ion, the

results will be diffi c u lt to i nterpret. The advant a ge of a stan-

dardized tes t , such as the o n e des c ribed , i s that e xper ienc e g a ined

i n assorted p r oj ects wi ll event ually be g rouped i n a useful f orma t

for f u t ure r eference.

Slaking of claystones and s h a l es can be c aused by swe ll ing of

clays. Tbe slak e durabi lit y tes t is n o t sui t ab le fo r swe l ling

materials as the lumps tend to bui ld p r ote ctive c l ay co a t ings.

Dur ab i lity p roblems associated wi th e xpans ive c l ay mi nera l s can be

predicted by stan d a r d methods of testin g fo r t he presence of swe ll-

ing clays an d measur e ment of swe l l p r essure. A conso li dometer

especially suited to t hi s purpose is the Ge oNor Swe lling app arat us**

in which disc- shape d rock samples, o r pulv e r i ze d an d e lutriated

samples, are pre c on so l idate d a n d then allowed t o s we ll un de r impo s e d

disp l acement c onstrai n ts ( Bje rrum, Brekke, et aI, 1963). F ree

swell of altered, hard rock samp les wa s meas u red witb s ufficient

precision very simp ly by Nasc imento ( 19 70) wi t h a jewe lled di al gauge .

The rock co r e s p ecime n st a n ds in a be ak er of wa t er on a po i n t con-

* ASTM Standar d Me t hods of Test C5 35-69 and C131-69 .

** So ld by the No r we g ian Geot echn i cal I nst i tute, Oslo.

29


tact. In several altered granites and gneisses Nascimento monitored,

swelling began almost immediately after the water was added and

essentially terminated after two to five hours.

A meaningful and potentially rewarding area of inquiry into the

weatherability of rocks considers the changing content of dissolved

solids in water percolating steadily through rock specimens. The

Bernaix radial permeameter, discussed later in connection with

fissuring, is suited to this approach.

INDEX TESTS FOR THE QUALITY OF THE ROCK MATERIAL

Other simple laboratory tests or quick field measurements can

serve as quantitative indices of rock quality and degree of weathering

and as basic components of applied classifications. Deere and Miller

(1966) studied the use of the Schmidt hammer which can be carried in

the field. Defects on the surface against which the hammer is

activated can give low readings unrelated to the rock material

quality, a problem which can be avoided by exercising care in pre-

paring the test surface. A more revealing measurement is provided

by any strength test, especially one which demands a small sample and

which can be done routinely on a large number of specimens. The

point load test (figure 2-3b) introduced by Franklin (1970) and Broch

and Franklin (1972) is one such method. Tests are conducted by

squeezing pieces of rock drill core diametrically between standard

steel cones until rupture. The point load index I is PjD2 where s P is the load at rupture and D is the diameter. The results are

affected by the value of D, but size correction charts given by

Broch and Franklin allow all results to be expressed in terms of a

standard size (50 mm is recommended). The point load index I s ,50

correlates fairly well with the uniaxial compressive strength

divided by 24. The test can also be applied to irregular chunks,

approximating 50 mm in size.

FISSURED ROCKS

Small cracks and fissures may be contained in apparently intact

rock specimens. As opposed to pores, which are three dimensional

(a)

(e) Figure 2-4. (a ) Scannmg electron microscope photo 01 a series of pores In

Berea sandstone. fWeinbrandt and Fatt.

19691. 102X. fb) Scanning electron

microscope photo o t another pore In

Berea sandstone ; (We inbrandt and FaIt,

1969). 1020X . (c) Polished section of

Mont Blanc granite. show ing Intense

ftssunng . (courtesy of Drs. M. Panet

and C. T ourenQ, Lab. des Ilonts et Chausees. PariS). l A X. (d) M,crophoto graph of a fissure fabriC reproduced In

a plastic fil m replica of a polished

diorite surface; (COurtesy of Dr. J.

Verdier. Coy ne and Beli ier , Par ies). SOX. fe) Fine fissures In granite, hi gh lighted

by dye and viewed in thin sect ion;

Huelgoal gran ite . Bri ttany; (courtesy of

(e) Drs. B. Schneider and J. Verd ier, Coyne and Bel1ier, PariS) .

(b)

(d)


(figures 2-4ab), fissures are short planar cracks of microscopic or

macroscopic size (figures 2-4 c,d,e). They occur as intercrystalline

cracks less than 1 micron to macroscopic (> Imm) in size, as inter-

grain cracks, and as multi-grain fractures. The presence of such

cracks as well as their significance in reducing the tensile strength

of brittle materials were appreciated by Hoek, Brace, McClintock and

Walsh, and others with regard to the Griffith theory of failure which

is based upon stress concentrations around such fissures. Habib and

Bernaix (1966) linked the degree of fissuring also with scale effects

in strength and deformation measurements, dispersion of results in

repeated measurements, and stress dependency of specimen permeability.

All of these effects were shown to be large in highly fissured rock

at low pressure and to disappear in non-fi$sured rocks, and in

fissured rocks at elevated pressure, within which the fissures have

closed. With respect to mechanical properties, it is the presence

of fissures more than any other aspect, wrote Habib and Bernaix, that

distinguishes rock from other solids. The French have held two

colloquia and an international symposium on rock fissures* and

correlations have emerged reinforcing Habib's belief that in fissured

rocks, mechanical properties are more closely dependent upon fissure

fabric than on mineral composition or texture.** It appears that

fissuring has a primary influence on static elastic modulus values,

hysteresis in load cycling, sound wave velocity, direct tensile

strength, resistivity and thermal conductivity of rock specimens.

The degree of fissuring in a rock reflects its history. Fissures

* The 1st and 2nd Colloquia on Fissuring of Rocks were published in special numbers of "Revue de l' Industrie Minerale" respectively 15 May, 1968 and 15 July, 1969. The Nancy Symposium held October 1971 was published by ISRM.

** In introducing the 2nd Colloquium on fissuring in rocks, Pierre Habib wrote: "One can now say that the properties of rocks are essentially those of their fissures. The mineral matrix has only a discreet role in the sense that if the rock is continuous it is always over endowed either in rigidity or in strength. To describe the fissuring of a rock is thus to define its present state and the study of its mechanical behavior is first of all the examination of the arrangement and development of fissures up to destruction."

Rock Classification

can b e gene r a ted b y chemi cal weac hering , unloading, heati ng and

c ooling , and most imp ortant l y b y loc ali z ed crack ing accomp a ny ing

deformation. Rocks likely to be fo un d in a fissured state are:

v o l cani c fl ow ro c ks; fo liated met amorph i cs, especi ally schistose

v a riet i e s; marble; pegmatites a nd porp h y ri t ic or hypidiomorphic

granit ic rocks ; g rani tes exhibiting cleavage; serpentine; chert and

sili c eous shales; a nd qua rtzi te.

The de gree of fis s ur i ng s h o u l d b e a basic compon e n t of any rock

c l assificat i o n s cheme. It c an be char acteriz e d through direct

observat ion , or mo r e simply throu gh index tests.

A pol ished s u rface will of ten enable promin e n t fissures to be

observed with t h e n aked eye. A han d lens, binocular microscope, or

bes t of all, an ore micros c ope al lows st udy of fissure distribution

in t h e p olished sect ion . Tourenq (196 9) di splayed f issures in

p ol ished surf a ces by prepari ng rep l i cas of the surface adapti n g

tech n i ques f or e lectron mic roscope spec i mens des cribe d by Bradley

(195 4) an d J acq uet and Me ncarel l i ( 1959 ) , (fi g u r e 2 -4d ) . Schneider

(196 7 ) used dyes: Fuschine ASA (basic), Victoria Blu e (basic ) and

Auramine J ( aci d) t o s h ow fi ne fis s u re detail in thin sect ions viewed

in tran s mitted ligh t with a pet rog rap h ic mi croscope, (figure 2--4e).

S tudy of fl u i d- fi l le d e pigeneti c i nclusions , e. g. in quartz, can

a llow r elat ive eva l ua t ion of dif f erent directio ns of fissuring,

(Verdier a n d Deicha, 19 71 ). These proc esses c a n be ted ious and for

pract ica l en g ineeri ng work it may be mo r e app ropriate t o characterize

th e deg r ee of f issuring imp lic itly.

Pa r a meters of fi s suri n g c an be de r i ved from pressure - volume

ch an g e cu r ves , shear a n d l ongitudina l wav e velocity meas u rement s ]

comp a ri s on of d i r ect a nd indire ct t ens i l e st rengths, and the rat io of

perme abilities in t ens i on and compression. Morl ier (1968), followi n g

wo rk of Wa l sh (1965) , s uggested c omput ation of the vo lume of f iss ures

--f i ssure p orosity f rom t he sh ap e of t h e p ressure - vo l ume change

curve (p versus tV ). As s h own in fig ure 2-5 , this curve is concave upwa rd , b e c omi n g a symptotic to a l i ne whose slope is defined by the

compres s ib i li ty of t h e r ock (k = 3( 1~2V ) ' The i n itial fis sure p orosi ty nf (o ) is est imate d b y the value of ~v at intercept of t he a symptote. The con cavi ty of t he cu r ve c a n also be interprete d to

33


Figure 2-5. Compressibility of fissured rock.

p

original rock volume = V

p

yield a fissure shape distribution function ("fissure spectrum").

Tourenq, Fourmaintraux, and Denis (1971) offered a second

approach based upon a comparison of actual and theoretical wave pro-

pagation velocities. A crystalline rock composed of given percen-

tages of stated minerals has theoretical elastic properties close to

the weighted average of the elastic properties of the components.

Table 2-3 gives values for Young's Modulus, Poisson's ratio, and

longitudinal and transverse wave velocities for the common rock

forming minerals. If a rock is fissured, measured properties will

be lower than the theoretical values calculated from Table 2-3. The

degree of fissuring is expressed in terms of a quality index, IQ,

defined as the ratio of measured to calculated longitudinal wave

velocities.

IQ V$I, measured

V$I, calculated x 100% (2)

Fissure porosity (n f ) drives the quality index downward linearly

approximately 15 times as fast as normal porosity (n ) (spherical p

pores). If one measures the total porosity n% (= np + n f ) as well

as IQ, figure 2-6 can be used to determine the value of n f . This

6V V

Rock Classification

TABLE 2-3

Average Elastjc Modulus and Velocity of Longitudinal Waves for Common Rock Forming Minerals *

quartz

ort hoc l ase

plagioc lase

biotite

c a l cite

mus c ov ite

amphibole

pyroxene

olivine

magneti t e

Young 's Mo dulus

E 5 ( 10 bar)

9.6

6. 7

8. 1

7.0

8. 1

7.9

12. 9

14 . 4

20 .0

23. 1

Poisson' s Rat io

v

0 . 08

0.27

0 .28

0.25

0.30

0 .25

0 .28

0.24

0.24

0 ~ 26

Longitudina l Velocity

V km / s e c

6 .0

5.7

6.3

5. 1

6.7

5.8

7.2

7.2

8.4

7 .4

Transver s e Veloci t y

Vt km/ sec

4. 1

3.3

3.5

3.0

3.4

3.4

4.0

4 .2

5.2

4.2

* From data of Ale xs androv , Be l ikov a n d Ryzova , a re f e rence c ited by Fourma intraux and Touren q ( 1970).

fi gure also shows t he re lative e ffect s of pores and f issures on t h e

rat io of measured to c alculated e lastic modu l us va l ues. A value of

n f ; 2% reduces t he i de a l elas tic mo du l us almost by half , whereas a

value of n ; 15% wou l d be req u ired to a chi eve this effe ct. p

If bot h t h e tra ns verse a nd l ongit udinal wa ve velocities a r e

measured, t he degre e of f i ss uri ng c a n be derived from t hei r ratio .

Fi ssure d rocks are not "ideal" materials and on e s hould not auto-

matically try to rep ort the r at i o Vt/V i n t e r ms f a I' dyn amic

Poisson ls ratioll value. I ns t ead , Tourenq et a l suggest Table 2-4.

A third method of evaluating t he de gree o f fis s uring i s base d

upon the ratio of s trengths i n di r ect an d indirect tension tests,

( Tourenq and De n i s, 1970) . Direct t ens i on tests can be performed by

bonding moment - free e nd pieces t o cy l indrica l r o ck spe cimens. The Goodman- Geul oglcal Eng.---4

35

36

~ o u ~

> "--.

iii o a.> E ~

>

Emeas./Eeale.

Rock Classification

Figure 2-6. Relative effect of fissures and pores on the longitudinal wave velocity and the modulus of elastic-ity; (Tourenq, Fourmaintreau, and Denis, 1971).

direct tensile strength, at, i.e. the average tensile stress at

failure, is greatly reduced by fissuring. An indirect tension test

(Brazilian test) can be obtained by compressing the opposite diameters

of a rock disc* causing a uniform state of tension across the vertical

diameter; the Brazilian tensile strength at B i.e., the tensile , stress at failure in a Brazilian test, is only slightly affected by

fissuring. Therefore, the ratio at / at,B is descriptive of the

degree of fissuring, as shown by Table 2-5 summarizing some data

presented by Tourenq and Denis. They recommend that the rock be

classed as: essentially non-fissured if at / at,~ > 0.8; very

fissured if at / at B < 0.2. , Bernai x (1969) developed an index of fissuring intensity based

upon a radial permeability test. Water introduced under pressure in

the center of a thick walled cylinder of rock, figure 2-7a, produces

tangential tension stress as it flows divergently towards the outer

circumference. Conversely, convergent flow produced by directing

water from the outer circumference to the inner produces a tangential

compression. Assuming that the flow net is not altered by stress

* In Tourenq and Denis' tests, the length to diameter ratio of the discs was unity.

Rock Classification

TABLE 24

Index to Degree of Fissuring According to the Ratio of Transverse to Longitudinal Wave Volocities

Description

< 0.6 non fissured

0 .6 to 0.7 f is s u red

> 0. 7 very fiss ured

TABLE 25

Tension Test Index to Fissuring

Data from To u r e n q a n d Denis (19 70 ) .

Ro ck Fissure length ( mrn)

Limest on e 0. 2

Limestone 1. 5

Gran i te 0.1

Granite 0.3

Gran ite 1.3

Gran i t e 2.5

Gr anite 1 to 10

Gr ani te 3 to 20

Basalt 0.1

Basal t 2 t o 10

t * t B ,

1.0

0.45

0 .93

0.7

0.50

0.34

0 . 14

0.07

0 . 9

0 . 15

* 0t = di re c t tensi on stre ng th; t B Brazilian tensile , st r e n gth .

37

38

150 100 12

(a)

Rock Classification

(/)

"-

'-4+-------0olitic limestone S=I

~ 10-8

10-10 2~0~~0~~2~0~~~~~6~0~~80~~100

P (bars)

(b)

Figure 2-7. Radial permeability test; redrawn from Habib and Bernaix (1966). (a) Radial permeameter (dimensions in mm). (b) Results for a porous and a fissured rock; S = K_,1K50

dependency of permeability*, Bernaix derived expressions for the

water pressure gradient dp/dr as a function of r, added this as a

body force to the equilibrium equations and solved for or and 08 ;

the value of 08 on the inner wall of the cylinder (r = R1 ) is

and

P 2(1-v) + (3 )

08 is tension for divergent flow where P is the water pressure

on the inner radius r = Rl ; (P(R2 ) 0)

08 is compression for convergent flow, where P is the water

pressure on the outer radius r = R2 ; (P(R l ) 0),

* In actual fact the permeability K is a function of and since is a function of r, K depends on rand dp/dr can not be solved as readily as in the paper. The problem is one of "coupled flow"; see Noorishad et al (1972).


For the conditions of Bernaix's tests, Rl = 0.6 cm, R2 = 3 cm

and Poisson's ratio 'J = 0.2 giving 08 = 1.53 P. Figure 2-7b shows

the variation of permeability K for an oolitic limestone and for

Malpasset gneiss as P was varied between 100 bars outside to 1 bar

inside. The permeability K was computed according to the relationship

K Q 2TT LP In (4)

wherein Q is the steady state flow rate (1 3 / t) and L is the length

of the cylinder. The permeability varied continuously over more than

3 orders of magnitude for the fissured gneiss whereas it remained

constant for the limestone. Bernaix recommends as an index of

fissuring to report the ratio (S) of permeability at 1 bar in diver-

gent flow (K_1 ) to permeability at 50 bars in convergent flow (K50 ).

A sampling of results with a number of rocks are summarized in

Table 2-6.

TABLE 2-6

Radial Permeability Test Index to Fissuring

Rock

Limestone

Limestone

Limestone

Grani te

Quartzite

Mica schist

Schist

Malpasset gneiss

Description of fissuring

porous, non fissured

porous, some fissuring

fissured

slightly microfissured

microfissured

fissured

highly fissured

highly fissured

* The failure originated on the left bank.

S

1

1.3

2.6

1.2

1.8

4.8

10 to 100

7 to 200 right bank 1 to 50,000 left bank*


DISCONTINUITIES

Rock masses invariably include numerous surfaces of real or

potential discontinuity. Though somewhat artificial, we will dis-

tinguish between discontinuities and fissures. Obviously, there is

a continuous distribution of discontinuity surfaces according to

length. However, fissures within a specimen are included in a sample

of the specimen, thus subject to meaningful inquiry in the laboratory.

Laboratory techniques for samples of larger surfaces of weakness are

developing but the results are seldom exportable to the field without

additional field observations and tests. Thus fissures can be con-

sidered as rock specimen features whereas discontinuities cannot.

A single discontinuity includes two mating surfaces and a space,

or filling. The term "joint" which has come to be used in engineer-

ing contexts for all or part of the family of discontinuities in rock

masses, is unfortunately potentially confusing for structural

engineers, who use the term joints to describe points of connection

in steel structures; in geological usage the term is applied only to

penetrative, repetitive discontinuities without appreciable shear

displacement. However, as the term joint is entrenched in its

engineering geology context, it will be retained here.

On a geological basis, we can distinguish extension and shear

joints, bedding, banding, contacts, cleavage, schistocity, foliation,

sheared zones and faults, as discussed in standard works in geology,

for example Leet and Judson (1971) and Price (1966). With reference

to mechanical and flow properties of a discontinuous rock mass, we

require considerably more information than the geological identifica-

tion. In particular, load-deformation and strength properties of

discontinuities (see Chapter 5) make specific reference to a number

of controlling quantities, including the parameters of the peak and

residual shear strength-variation with normal stress, the initial

angle of dilatancy, the normal pressure required to prevent all

dilatancy, the maximum amount a joint can close, the peak and residual

shear displacement, the tensile strength, and the normal and shear

stiffness. Though no rational formula exists for extracting the

explicit joint parameters required for an analysis from field observa-

tions, careful and detailed descriptions of the many encounters in

Rock Classification

out c r ops, excava tions , and in the core box a l low the who le syst e m of

disconti nuities in a ny project to be di vided int o a r elatively smal l

number of types. Us ually the fiel d des cript ion wil l permit reason-

able estima tes fo r certain o f the q uan t ities me ntioned and their

contribution t o the rock mass char act e rist ics. The j o int system

prope rt ies are de r ive d from observable feature s o f: 1 ) the discon-

tinuity surfaces ; 2) t he bl ocks they de fi ne t hrough t hei r repetiti on

an d intersection; 3 ) an d the p roperties o f t h e space between t he

blo cks. Most of t h e se feat ure s can b e de s cribe d deterministically ,

but are better expressed stat isti cal ly through dis t ribution curves

and numer i cal in di ces.

41

Prop erties of t he joi n t s urfaces thems elves i nclude orientation ,

exten t, p lanarit y , r ou ghness and waviness, an d t he strengt h of wall

rock a s peri ties . J o in t systems usual l y di s p l ay several preferred

ori e ntat i ons and t hi s a s pec t a l on e res ult s in wholly d i fferent classes

of ro ck masses. Th is sub j e c t l ends i t se l f t o st at ist ical and graphi-

cal treatment by me an s of s te re og r aphic proje c tion , which wil l be

elabora t ed i n Chapter 3. Join t "e xten t ", mean i ng the t otal area or

length, us ually cann o t b e me asured dire c t ly in t h e f ield; however it

can b e estimated occas ionally from aeria l photos (see Chapter 4).

Roughness a n d wavines s, whi ch influ ence t he f ri ct ion a ngl e s ] di l atancy,

an d peak s hear (Pa t ton 1966 , Goodman and Dubois, 1971) refer to the

local departures from p l anarity a t s ma ll and large scales respe ctively

(fig ure 2-8) . Th e most c onvenie n t roughness measure f or rock mechan-

ics p u rposes i s i n terms of the l oca l angles with respect t o the mean

plane through all the hills a nd val l ey s of a j oint surface . Most

joints can b e r e pre sent e d qui te wel l b y p l anes. P a t ton (1966)

meas ure d r oughness ang les f rom e dge views of s urfaces in outcrops and

cli f f faces ; t h e required data can be ob t a ine d from photographs as

di s cussed i n Ch ap ter 4. When the joi n t surface i t sel f is expos ed in

outcrop , repe a t ed measur ements o f d i p wi ll gene rate a scattered dis-

tr i bution of va lues, whose st an dar d devi a t ion o r mean departure may

be an estimate of t he me an r oughness a ng le, a s discuss e d in Chapter

5. The roughnes s ang l es i ncrease joint s h e ar strength at l ow normal

pres sure, but at hi ghe r norma l p ressure, t he strength o f t he wal l

ro ck asperit ies c ont rols t he s he ar strength of the joints. A good

(a)

(c)

(d)

Figure 2-8. (a) A very rough fracture surface in granite; note the perfect mating of the joint blocks across the tracture. (b) A very rough bedding plane in limestone; voids between beds resulted trom oversliding of asperities accompanying mass movement downslope. (c) A rough bedding plane surface; the roughness is tormed by ripple marks preserved from the depositional surface. Photo by Dennis Lachel, (courtesy of the Corps of Engineers). (d) A rough joint surface; the roughness is created by the intersections of the joint with bedding; erosion has accentuated the rel ief. The rock is an argillite. Photo by Dennis Lachel, (courtesy of the Corps of Engineers). (e) A smooth surface whose mean plane parallels bedding.

methods of geological engineering in discontinuous rocks

Documents